
Building an AI Chief of Staff: Ambient by Lawrence Coburn
with Lawrence Coburn, Ambient
Building an AI Chief of Staff: Ambient by Lawrence Coburn
Show Notes
Lawrence Coburn has been through the full founder cycle - twice. Double Dutch raised $93 million, expanded to offices around the world, hit product-market fit faster than the team could handle, then got acquired by a competitor backed by Vista Private Equity. After that exit, Lawrence built Twine, a remote-work networking product that was dependent on pandemic conditions in a way that made him uncomfortable. When ChatGPT launched, he recognized the wave immediately. He had been at the start of the internet, mobile, and social - this one felt bigger.
Ambient is the product Lawrence wished he had when he was drowning at Double Dutch. His board told him to hire a chief of staff. He did not know what that was. Then he met Matt - Harvard Business School, ex-Bain, able to make decisions with incomplete information - and his bandwidth effectively doubled. Ambient is the AI version of that: a chief of staff for the 4.2 million companies that do not have one. Three workflows, a team of three, $1.5 million raised, and a founding insight that has held up through 400 interviews: the best chiefs of staff are extraordinarily talented people who have completely checked their ego at the door.
What Ambient Actually Does
Ambient identified three workflows that a chief of staff performs that AI is already good at. The first is daily prep: before each meeting, an AI agent researches everyone on the calendar - their background, recent activity, shared connections, investment thesis (if it is a VC), any red flags - and delivers a briefing dossier to the CEO's inbox. The second is rhythm of business: secure, structured note-taking for executive staff meetings that captures who is doing what, flags misalignment, and drives follow-through. The third is key initiative tracking: AI agents monitor Slack, email, transcripts, and other connected systems to surface red/yellow/green status on the make-or-break projects every CEO should be watching.
A standout feature is fundraising mode. When an AI qualifier agent detects a VC on the calendar, the briefing shifts: it surfaces icebreakers and things in common, similar portfolio investments, publicly stated thesis, any founder squabbles or lawsuits, and people you know in common. Lawrence used it the morning of an investor call and discovered the investor was a competitive America's Cup sailor - the same sport his co-founder competes in. That kind of personal connection in the first 180 seconds of a pitch can be the difference between a cold deck and a closed round.
5 Frameworks from This Episode
1. The 1.8x Throughput Model
- A truly talented human chief of staff gave Lawrence roughly 1.8x his effective bandwidth - not just more hours, but better decisions, better-prepared meetings, and better execution follow-through
- AI can realistically target the same ceiling for CEOs; engineers can get to 5x because their workflows are more repeatable and measurable
- The 1.8x is not just about time recovered - it includes quality of decisions: a CEO who understands all relevant options before choosing is more valuable than one who moves faster but makes worse calls
- Lawrence's former chief of staff framing: "This feels directionally correct." The AI era should raise the bar beyond directionally correct to high-quality decisions with full context
2. The Calendar-as-Lens Framework
- The CEO's calendar is the most honest signal of what the company actually prioritizes - more honest than strategy docs, OKRs, or mission statements
- If it is not on the calendar, it is not happening - which means the calendar is a searchable history of every meaningful relationship, decision, and initiative
- Ambient's core bet: earning access to the calendar unlocks the intelligence layer - briefings, relationship tracking, initiative monitoring - that everything else depends on
- Future extension: calendar-to-CRM jumps (you haven't spoken to this person in 6 months; their portfolio company just sold; it may be time to reconnect)
3. Push Not Pull Intelligence
- The old paradigm: insights live in a dashboard that the CEO has to log into, remember to check, and learn to navigate
- The new paradigm: intelligence finds the CEO where they already are - the inbox - without requiring login, onboarding, or behavioral change
- Lawrence's observation: senior CEOs at later-stage companies are not on Product Hunt testing new tools; to reach them, you have to come to them
- OpenAI's Pulse product is experimenting with the same principle at consumer scale: how do you surface what is interesting before the user asks?
- Ambient's delivery mechanism - daily briefing via email - is a product decision, not a technical constraint
4. The Narrow ICP Moat
- In the age of generative AI, there are no technical moats - every startup has access to the same foundation models
- The only durable protection is going narrower on ICP than competitors are willing to go, and knowing that ICP better than anyone else
- Lawrence knows the founder ICP from the inside - he is one, has been one through multiple cycles, and interviewed 400 chiefs of staff to map the workflows
- Narrow ICP also unlocks depth: fundraising mode, exec staff meeting note-taking, initiative tracking - features that are irrelevant to a generic knowledge worker but essential to a CEO
5. The High-Quality Decision Standard
- A CEO makes hundreds of decisions per month; one or two could kill the company - and you do not know which ones those are in advance
- The chief of staff concept of a "high-quality decision" means: even if the decision turns out to be wrong, you properly gathered the information and weighed both sides
- Because you cannot predict which decision is the kill shot, every decision deserves the same rigor - which is exactly what a chief of staff (or AI agent) can provide at scale
- The AI chief of staff does not improve the CEO's instincts - it improves the information environment those instincts operate in
Founder Experiment: Audit Your Calendar for Hidden Intelligence
Step 1 - Export your last 90 days of calendar events. Pull every meeting with an external person - investor, customer, partner, candidate. List them in a spreadsheet. This is your company's actual relationship map, more honest than your CRM.
Step 2 - Identify the meetings where you were under-prepared. Which calls did you walk into without knowing the other person's background? Which investor meetings did you not know their portfolio or thesis? Which customer calls did you not have context on their recent activity? These are the gaps an AI briefing system closes.
Step 3 - Map your make-or-break initiatives against your calendar time. Is your calendar actually reflecting your stated priorities? If your top three strategic initiatives are not showing up as regular blocks, they are not really your top three. The calendar does not lie.
Step 4 - Run a relationship decay audit. Who is on your calendar from 6-12 months ago that you have not touched since? Which of those relationships matter to your current priorities? This is the type of proactive reconnection an AI chief of staff surfaces automatically.
Step 5 - Before your next 5 most important meetings, spend 10 minutes on manual prep using LinkedIn, recent news, and shared connections. Document what you find and how it changed the quality of the meeting. This is the baseline of what Ambient automates - knowing what it produces manually tells you exactly how much value the automation unlocks.
Glossary
Tools & Resources Mentioned
Q&A
What is the founding story behind Ambient?
Lawrence was the CEO of Double Dutch during its rocket-ship phase - 93 million raised, offices worldwide, on a plane two out of every three days. His board saw him drowning and recommended he hire a chief of staff. He didn't know what that was. He brought in Matt, who had been at HBS and Bain, and the experience was transformative: better-prepared pitches, clearer positioning, dramatically more bandwidth. In the age of AI, Lawrence decided to build the version of that for the 4.2 million companies that don't have a chief of staff and probably should.
What is the difference between an executive assistant and a chief of staff?
The primary tell is timeframe. EAs are focused on the day-to-day: calendar management, inbox, scheduling, sometimes week-ahead prep. Chiefs of staff operate on a longer horizon - M&A, VC pitch preparation, data rooms, strategic planning. Lawrence's two diagnostic questions: does this person run your strategic planning process, and do they create the agenda for your executive staff meetings? If the answer to both is no, you have an EA with a chief of staff title, not a true chief of staff. Both roles benefit enormously from full context access - that's the common thread with AI agents.
What is the 1.8x throughput target and where does the ceiling come from?
Lawrence's experience with his own chief of staff Matt gave him roughly 80% more effective bandwidth - essentially 1.8x his solo output. That is the ceiling Ambient is targeting for AI. It is lower than what engineers can achieve (Lawrence's CTO reports 5x using Cursor) because CEO work is less repeatable and involves relationship judgment and institutional knowledge that resists automation. But 1.8x is not just time - it includes better-quality decisions made with full context, which Lawrence argues is worth as much as the time recovered.
Why does Lawrence deliver Ambient's intelligence via email instead of a dashboard?
Because senior CEOs do not learn new tools. At early-stage companies, founders are builders who will try anything. At later-stage companies, the CEO is in their inbox, in Google Docs, maybe occasionally in the CRM. They are not on Product Hunt. Ambient's philosophy is that AI intelligence has to meet the CEO where they already are - if you require them to log into a new tool, you've already lost. The daily briefing delivered to the inbox is a product design decision, not a technical compromise. Push, not pull.
How does Ambient's fundraising mode work in practice?
An AI qualifier agent reads every calendar event and classifies the meeting type. When it detects a VC, it triggers special treatment: the briefing dossier is restructured to prioritize icebreakers and personal connections (Lawrence discovered his investor was an America's Cup sailor - his co-founder competes in the same sport), similar portfolio investments, the investor's publicly stated thesis, any red flags or founder disputes, and mutual connections. Lawrence's rule: in a 30-minute pitch, you have about 180 seconds to make a personal connection. The briefing controls everything you can control.
What are Lawrence's two tells for identifying a true chief of staff?
First: do they run the strategic planning process? Second: do they create the agenda for the executive staff meeting? If the answer to both is no - if someone else owns strategic planning, or if the chief of staff is just taking notes - then regardless of the title, the role is more EA than chief of staff. Lawrence uses these questions when interviewing candidates and when evaluating whether a company genuinely has chief of staff capability or just the title.
What surprised Lawrence most from interviewing 400 chiefs of staff?
The scale of talent combined with the absence of ego. He met a chief of staff at a major Walgreens division whose previous role was running a $10 billion organization. These are people who could be running their own companies. But the best chiefs of staff have completely removed themselves from the equation: their name is not in the press release, they have no budget or headcount to lobby for, and their only metric of success is the success of their principal and the company. Lawrence calls it one of the purest roles in business.
Why does Lawrence believe senior EAs and chiefs of staff will be the last people standing in the AI economy?
Because they can flex up. Every time a lower-skill task gets automated - note-taking, scheduling, inbox management - a talented chief of staff or EA moves to a higher-value task. If you take everything away, they are still the trusted person in the foxhole with the executive: the one who has context no system has, the one who can exercise judgment no model has. Lawrence went public with this view because he thinks the conventional wisdom about AI displacing these roles is wrong - adaptable generalists with deep executive access are exactly the profile that is hardest to replace.
What is Lawrence's take on the current San Francisco energy and the AI wave?
He says it's back - and not in a subtle way. Real estate in his Mission district neighborhood is moving fast (a house listed at $1.8M sold for $3.2M). People who moved to Miami during the pandemic have returned. Anyone working on AI is flocking to the city. Lawrence has been through the internet wave, the mobile wave, and the social wave. His assessment of this one: it is the big one. The accessibility is what makes it different - 800 million ChatGPT users means the software reached everyone, not just technologists. And the production side, English as a programming language, means anyone can now build what was previously only possible for engineering teams.